INDEX

Explanations

dates written in a specific format (year)

oai_token-act-pair · gpt-3.5-turbo

repeated instances of the sequence "yr" in varying contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Ou

-0.86

Ic

-0.68

 Papua

-0.68

 Fiesta

-0.66

 Shack

-0.64

 Labyrinth

-0.64

iership

-0.64

 Samoa

-0.63

leeve

-0.63

ļé

-0.63

POSITIVE LOGITS

rha

1.31

interstitial

0.90

rh

0.90

rr

0.89

andom

0.83

azines

0.79

haps

0.78

rocal

0.78

annis

0.76

umph

0.76

Activations Density 0.015%

dates written in a specific format (year)

repeated instances of the sequence "yr" in varying contexts

No Comments

No Known Activations